| Field | Details |
|---|---|
| Team Name | Team Operionix |
| Hackathon Name | MSME Idea Hackathon 6.0 |
| Project Title | Operionix |
| Tagline | Unified Autonomous Multi-Agent and Edge IoT Platform for MSME Operations, Sustainability, and Credit Underwriting |
| Team Motto | Innovate, Build, Impact |
| Team Members | Godfrey and 5 other team members |
| System Architect | Godfrey |
| Contact Email | godfreytr.prof@gmail.com |
| Mentor | Prof.Jasmin R |
| Selected Theme | Industry 4.0 and 5.0 |
| Selected Idea Sector | Miscellaneous Sector (Environment, Forests, Water & Sanitation; Foods, Beverages, FMCG, Consumer Goods; Infrastructure, Construction, Housing; IT, ITES, Electronics, White Goods, Telecommunication; Metals, Engineering, Machinery, Automation and Transportation, Automotive, E Vehicles, Railways, Aviation, UAV and any other sub-sector) |
Micro, Small, and Medium Enterprises (MSMEs) face severe constraints when trying to digitize, automate, and transition to sustainable business practices:
- Siloed & Prohibitive Automation: Existing software (ERPs, GST tools) and hardware (robotics, EV telematics) are highly fragmented. MSMEs cannot afford to purchase and integrate multiple standalone platforms.
- Severe Credit Gaps & Capital Lockups: Underwriting remains slow and asset-collateral dependent. Traditional credit scoring fails to reward green practices or factor in complex government energy subsidies. Furthermore, delayed supplier GST filings lock up critical Input Tax Credit (ITC) working capital, which is fatal to cash-strapped MSMEs.
- Unbankable EV Fleets & Telematics Gaps: MSMEs transitioning to electric vehicles (EVs) for logistics face high capital depreciation due to unpredictable battery degradation under harsh Indian climates. Lenders refuse to refinance used EVs without battery health metrics. Critically, low-cost Indian e-rickshaws and e-loaders utilize cheap motor controllers/BMS units with no CAN bus or OBD-II telemetry output.
- Hazardous, Dusty Production Floors: Small manufacturing plants rely on manual quality control, leading to high reject rates (up to 7%) in export markets, while workers remain exposed to mechanical hazards. Standard optical vision systems (RGB-D depth cameras) fail quickly in these environments due to heavy dust, vibration, and smoke blocking the lenses.
- Target Users: MSME business owners, factory managers, logistics fleet operators, and compliance officers.
- Industry: Small-Scale Manufacturing, Last-Mile EV Logistics, Renewable Energy (Solar CAPEX), and FinTech/Banking (cooperative and public sector banks, rural lenders, NBFCs).
- Community: Factory floor workers (occupational hazards), local supply chain ecosystems (working capital locks), and local communities (air pollution, and toxic battery waste due to premature disposal).
Addressing this problem is critical because:
- Economic Viability: Releasing locked tax credits directly boosts the working capital of thin-margin MSMEs.
- Global Competitiveness: Automating manual quality control reduces reject rates from 7% to under 1%, allowing MSMEs to compete in export markets.
- Sustainability Transition: Accelerating solar integration and EV adoption relies on unlocking green credit subventions and battery refinancing options.
- Occupational Health & Safety: Continuous proximity tracking around automated equipment prevents catastrophic worker injuries.
Operionix is a unified, modular "Agentic and IoT Operating System" designed as a low-cost integration middleware. It connects an MSME's compliance records, supply chain negotiations, vehicle fleet telemetry, and factory floor sorting under a single stateful intelligence layer. Instead of a monolithic, high-cost platform, Operionix is designed as a decoupled, modular hub where MSMEs deploy only the services they need, utilizing a unified PostgreSQL/TimescaleDB ledger to feed de-risked underwriting metrics to banking partners:
- EcoVest Subsystem (Green Credit & ESG Auditing): Evaluates solar installation sizing, MNRE subsidies, and ESG compliance, generating a "Bankable Operational & ESG Audit Report" that integrates with SIDBI green lending guidelines to secure interest subventions.
- SupplyAgent Subsystem (GST & Sourcing Swarm): Runs stateful LangGraph agents (Negotiator, Auditor, Eco-Logistics) to automate supplier catalog searches, negotiate prices, and match invoices with GSTR-2B records to resolve tax credit locks via automated WhatsApp alerts.
- VoltSentry Subsystem (EV Fleet Battery Diagnostics): Uses a hybrid telemetry capture method—either a low-cost OBD-II CAN logger OR the Operionix Retrofit Sensor Board (an inline shunt-based current, voltage, and temperature sensor for non-CAN battery packs)—combined with cloud-based LSTM networks to estimate battery State of Health (SoH) and Remaining Useful Life (RUL), providing verified bank-grade battery health scorecards.
- CobotSight Subsystem (Human-Safe Vision-Guided Cobotics): Runs YOLOv10-nano active safety loops on Jetson Orin Nano edge computers. To prevent optical sensor blindness in dusty workshops, it implements a sensor-fusion fallback combining an Intel RealSense D435 camera (in a pressurized, air-purged IP65 enclosure) with industrial ultrasonic proximity sensor arrays.
🔍 Click to expand 3.1.1 EcoVest Module (Green Credit & ESG Auditing)
- Operational Flow: Ingests factory building plans, roof area, and historical electricity bills. A dedicated Sizing Agent runs optimization algorithms (e.g., PVWatts API or mathematical sizing models) to calculate optimal solar array capacity (kWp) and payback period. A Subsidy Agent pre-fills the MNRE Component-C application forms. The ESG Auditor Agent pulls tax ledger compliance logs and carbon metrics to score the company and signs the "Bankable ESG score card".
- RAG-Enabled Underwriting: Utilizing the
pgvectorstore loaded with SIDBI guidelines and regional state tariff rates, the system builds an explainable underwriting profile. EcoVest compares the MSME's ESG score against green capital discount benchmarks, generating a signed audit packet. This packet is output in a structured XML format compatible with the RBI Account Aggregator (AA) rails for seamless bank ingestion.
⚙️ Click to expand 3.1.2 SupplyAgent Module (GST & Sourcing Swarm)
- LangGraph Sourcing Swarm: Integrates three specialized agents:
- Negotiator Agent: Executes multi-turn game-theoretic bidding models (setting reservation prices and concession curves) with supplier API endpoints.
- GST Auditor Agent: Automatically matches GSTR-2B invoices against MSME purchase ledgers using fuzzy-string alignment and transaction-level checksums to identify delays or omissions.
- Eco-Logistics Agent: Uses GHG Scope 3 logistics routing to calculate estimated carbon footprints based on transport distance, payload, and fuel type.
- Autonomous Escalation & Twilio Loop: When GSTR-2B discrepancies are detected, the GST Auditor generates a pre-formatted legal notification citing invoice details and payment links. It triggers Twilio to send this notice directly to the delinquent supplier’s WhatsApp. If resolved, it writes to the Supabase ledger, updating the MSME’s cash-velocity profile.
🔋 Click to expand 3.1.3 VoltSentry Module (EV Fleet Battery Diagnostics)
-
Edge EKF Modeling: An onboard ESP32-S3 microcontroller captures battery pack data (voltage
$V$ , current$I$ , and temperature$T$ ) from either the OBD CAN line or the analog smart shunt retrofit. Running a discrete-time Extended Kalman Filter (EKF), the ESP32 estimates internal resistance ($R_0$ ) and State of Charge (SoC) on the fly, storing anomaly snapshots locally. - Cloud LSTM Pipeline: A local MQTT client streams anomaly triggers and 10-minute battery performance averages to the cloud backend. A pre-trained LSTM neural network evaluates these profiles to predict remaining cycle life, State of Health (SoH), and thermal runaway trajectories. These metrics compile into a signed battery health scorecard, serving as bankable valuation collateral for EV refinancing.
🤖 Click to expand 3.1.4 CobotSight Module (Human-Safe Vision-Guided Cobotics)
- Jetson YOLOv10 Active Loop: The Jetson Orin Nano runs a TensorRT-quantized YOLOv10-nano model on RGB-D camera feeds to detect operators and classify workplace safety boundaries.
- Proximity Sensor-Fusion: Merges camera bounding boxes with depth values and backup ultrasonic sensor telemetry.
- Dynamic Velocity & Interrupt Control: A ROS2 node monitors Euclidean distances between workers and the igus Delta arm:
- Safe Zone (>1.2m): ROS2/MoveIt2 operates the igus delta arm at full path velocity.
- Warning Zone (0.5m - 1.2m): Velocity is scaled down linearly to prevent sudden momentum.
- Inner Danger Zone (<0.5m): Sends a low-level interrupt to the TB6600 stepper drivers to lock the motors immediately.
- Multi-Agent Stateful Orchestration: Interactive workflows built on LangGraph that cross-optimize financial, procurement, and logistics tasks.
- Autonomous GST Discrepancy Resolution: Fuzzy string/numerical invoice auditing with automated Twilio-based WhatsApp notifications to delinquent suppliers.
- Retrofittable EV Battery Diagnostics: Adaptive CAN logging and shunt-based analog sensor telemetry evaluated via Extended Kalman Filtering (EKF) on edge ESP32 devices combined with Cloud LSTM models, generating bankable health ratings for low-cost fleets.
- Sensor-Fused Zero-Cage Dynamic Safety: Dynamic distance monitoring combining depth-sensing cameras and backup ultrasonic sensors to decelerate or halt robotic arms in dusty/hazardous industrial environments.
What makes our solution unique?
- Novel Approach: Coordinates physical edge telematics (VoltSentry, CobotSight) and digital finance (EcoVest, SupplyAgent) under one cooperative multi-agent system, linking operational data directly to credit eligibility.
- Competitive Advantage: Lowering the entry barrier for high-end automation from ₹15 lakhs to under ₹1.5 lakhs using retrofitted sensors, low-cost edge controllers, and open-source models, while simultaneously unlocking cheaper capital (50–150 bps discount) for sustainable practices.
- Technical Innovation: Edge-quantized YOLOv10-nano active safety zones, local-to-cloud telemetry filtering (reducing cellular data costs by 80%), and stateful multi-agent pipelines (LangGraph + pgvector RAG) mapping policy documents directly to business logic.
Operionix introduces several pioneering architectural and operational breakthroughs that differentiate it from conventional enterprise middleware and industrial IoT platforms:
-
Closed-Loop Cyber-Physical & Financial Fusion: Traditional systems keep physical automation (IoT, cobots) completely separate from corporate accounting and banking. Operionix establishes a direct, real-time feedback loop. For example, a safety event on the factory floor (detected by CobotSight) or battery telemetry (from VoltSentry) directly influences the MSME's operational integrity score, which dynamically updates their green credit rating and interest subvention margins at participating lenders.
-
Decoupled Edge-to-Cloud Stateful Orchestration: Instead of a heavy cloud-only or edge-only system, Operionix uses a hybrid architecture:
- At the Edge: Ultra-low-latency ESP32-S3 controllers run discrete-time Extended Kalman Filtering (EKF), and Jetson boards run YOLOv10-nano active safety loops.
- In the Cloud: Stateful LangGraph agent swarms analyze long-term patterns, perform fuzzy invoice auditing, and run LSTM networks for predictive battery health forecasting. This drastically reduces telemetry cellular bandwidth costs (by 80%) while ensuring microsecond-level local safety actions.
-
Sub-₹1.5 Lakh Retrofitted Automation Middleware: Instead of demanding that MSMEs replace legacy machinery and fleet vehicles (which would cost over ₹15 lakhs), Operionix is designed as a retrofittable middleware:
- For Fleet Vehicles: The custom Smart Shunt Retrofit Board captures raw analog voltages and currents from cheap, non-CAN bus batteries.
- For Robotic Safety: Pressurized, air-purged IP65 housings and backup ultrasonic sensors turn standard depth cameras into industrial-grade safety zones.
-
Consent-Driven Bank Integration via Account Aggregator (AA) Rails: Rather than locking credit evaluations behind proprietary banking portals, Operionix compiles multi-agent compliance audits, carbon indices, and ESG footprints into a standardized, digitally-signed XML format fully compatible with the RBI's Account Aggregator (AA) framework. MSMEs maintain full ownership of their data and can securely consent to share their green credit scorecards with any national lender.
| Category | Technology |
|---|---|
| Frontend | Next.js 15 (App Router), TypeScript, Tailwind CSS, shadcn/ui, Recharts |
| Backend | FastAPI, LangGraph, LangChain, LiteLLM API Gateway, Python |
| Database | PostgreSQL (Supabase), TimescaleDB (HYPERTABLE), pgvector |
| AI/ML | Llama-3-70B (Groq API), Claude 3.5 Sonnet, YOLOv10-nano (TensorRT), LSTM Models |
| Cloud | Supabase Cloud, Groq/Anthropic hosting, Cloud VMs for AI models |
| APIs | Setu/Sandbox APIs (GSTN & Account Aggregator), Twilio WhatsApp API, Microsoft Presidio (PII scrubbing) |
| DevOps | FreeRTOS (C/C++ for ESP32/ARM), ROS2, MoveIt2, MQTT broker, AES-256 encryption |
%%{init: {
'theme': 'base',
'themeVariables': {
'primaryColor': '#E2E8F0',
'primaryTextColor': '#1E293B',
'primaryBorderColor': '#94A3B8',
'lineColor': '#475569',
'secondaryColor': '#F1F5F9',
'tertiaryColor': '#E2E8F0',
'mainBkg': '#FFFFFF',
'nodeBorder': '#94A3B8',
'clusterBkg': '#F8FAFC',
'clusterBorder': '#CBD5E1',
'titleColor': '#0F172A',
'edgeLabelBackground':'#FFFFFF',
'actorBorder':'#94A3B8',
'actorBkg':'#E2E8F0',
'actorTextColor':'#0F172A',
'actorLineColor':'#475569'
}
}}%%
flowchart TB
subgraph UI[User Interface & Presentation Layer]
dash[Next.js 15 Web Dashboard]
echarts[Apache ECharts Visuals]
leaflet[Leaflet.js Map Views]
dash --- echarts
dash --- leaflet
end
subgraph GW[Gateway & Agentic Orchestration Layer]
fastapi[FastAPI System Gateway]
litellm[LiteLLM API Gateway]
subgraph LangGraph[LangGraph Stateful Agent Swarm]
ecovest[EcoVest ESG & Sizing Agent]
supply[SupplyAgent Sourcing & GST Agent]
volt[VoltSentry Predictive Fleet Agent]
cobot[CobotSight Edge Safety Agent]
end
fastapi <--> litellm
fastapi <--> LangGraph
end
subgraph DB[Database & Persistence Core]
postgres[(Supabase PostgreSQL Ledger)]
timescale[(TimescaleDB Hypertable)]
pgvector[(pgvector Document Store)]
end
subgraph Edge[Edge IoT & Hardware Layer]
subgraph VoltSentry[VoltSentry Hardware]
esp32[ESP32-S3 Microcontroller]
can[MCP2515 CAN Bus Controller]
shunt[Smart Shunt Retrofit Kit]
esp32 --- can
esp32 --- shunt
end
subgraph CobotSight[CobotSight Workstation]
jetson[NVIDIA Jetson Orin Nano]
realsense[Intel RealSense D435 Camera]
ultrasonic[Ultrasonic Proximity Array]
ros2[ROS2 MoveIt2 Controller]
jetson --- realsense
jetson --- ultrasonic
jetson --- ros2
end
end
subgraph Ext[External Services & APIs]
twilio[Twilio WhatsApp API]
setu[Setu GSTN & Account Aggregator]
llm[Foundation Models: Groq & Anthropic]
end
%% Connections
UI <-->|REST / WebSockets| fastapi
fastapi <--> DB
LangGraph <--> DB
%% IoT connections
esp32 -->|MQTT Telemetry| timescale
jetson -->|Safety Actions & Visuals| postgres
%% External API connections
litellm <--> llm
supply -->|Escalation Alerts| twilio
ecovest & supply -->|Filing Checks & AA Rails| setu
%%{init: {
'theme': 'base',
'themeVariables': {
'primaryColor': '#E2E8F0',
'primaryTextColor': '#1E293B',
'primaryBorderColor': '#94A3B8',
'lineColor': '#475569',
'secondaryColor': '#F1F5F9',
'tertiaryColor': '#E2E8F0',
'mainBkg': '#FFFFFF',
'nodeBorder': '#94A3B8',
'clusterBkg': '#F8FAFC',
'clusterBorder': '#CBD5E1',
'titleColor': '#0F172A',
'edgeLabelBackground':'#FFFFFF',
'actorBorder':'#94A3B8',
'actorBkg':'#E2E8F0',
'actorTextColor':'#0F172A',
'actorLineColor':'#475569'
}
}}%%
flowchart TD
subgraph Data Capture & Edge Layer
A1[MSME GST Invoices & Bank Logs] --> B1[OCR & Data Extraction Engine]
A2[EV Telematics: CAN Bus / Retrofit Shunt] --> B2[VoltSentry Edge EKF Controller]
A3[Cobot Sensors: Depth Camera + Proximity Arrays] --> B3[CobotSight Edge YOLOv10 & Sensor-Fusion]
end
subgraph Data & Storage Core
B1 -->|DPDP Consent & Presidio Scrubbing| C[Unified Supabase PostgreSQL Ledger]
B2 -->|MQTT Streams| D[TimescaleDB Time-Series Data Store]
B3 --> C
D --> E[Vector DB: pgvector RAG Store]
C --> E
end
subgraph Stateful Agentic Orchestrator [LangGraph Swarm]
E <--> F1[Compliance & GST Auditor Agent]
E <--> F2[Sourcing & Bidding Negotiator Agent]
E <--> F3[Lending Underwriter Agent]
E <--> F4[Predictive Fleet Maintenance Agent]
end
subgraph Execution & Output
F1 -->|WhatsApp API| G1[Delinquent Supplier Alerts]
F2 -->|API Calls| G2[Vendor Catalog Bidding Nodes]
F3 -->|XAI Memos| G3[SIDBI-Compatible ESG Underwriting Scorecard]
F4 -->|ROS2 Control| G4[Robotic Arm Trajectory Controllers]
end
Start (MSME Inquiry / IoT Trigger)
↓
User Input
(Operational logs, query, or sensor feed)
↓
Data Processing
(Edge filtering, EKF, or OCR processing)
↓
Core Logic / AI
(LangGraph orchestrator, YOLO, or LSTM models)
↓
Generate Results
(GST reports, safety halts, or health scorecards)
↓
Display Output
(Next.js dashboards, ROS2 active brakes)
↓
End
sequenceDiagram
participant MSME as MSME Operator
participant NEGA as Negotiator Agent
participant VEND as Supplier Endpoint
participant LEDG as Purchase Ledger
participant COMP as Compliance Agent
participant GSTN as GSTN GSTR-2B API
participant TWIL as Twilio WhatsApp API
participant UNDR as Underwriting Agent
participant BANK as Bank Credit Portal
MSME->>NEGA: Query raw materials (best price & eco-score)
NEGA->>VEND: Query catalogs & negotiate bid
VEND->>NEGA: Return negotiated price & delivery
NEGA->>LEDG: Write invoice & purchase ledger
LEDG->>COMP: Trigger GST reconciliation scan
COMP->>GSTN: Pull GSTR-2B filing status
GSTN->>COMP: Mismatch detected (delinquent supplier)
COMP->>TWIL: Trigger supplier reminder pipeline
TWIL->>VEND: Send WhatsApp reminder/payment alert
COMP->>UNDR: Report resolved mismatches & clean ledger
UNDR->>BANK: Update real-time cash velocity & adjust credit limit
- Manufacturing & Sourcing MSMEs: Small-scale industrial workshops seeking affordable automation (CobotSight) and automated supply chains.
- Last-Mile EV Logistics Fleets: Transport operators looking to de-risk battery degradation and secure low-interest EV refinancing.
- Financial Institutions (Rural & Cooperative Banks): Lenders requiring automated green underwriting tools and verified collateral scorecards.
- Democratized Automation: Lowers entry barriers for automated safety & sorting from ₹15 lakhs to under ₹1.5 lakhs using open hardware and open-source models.
- Enhanced Worker Safety: Mitigates the risk of catastrophic fleet battery fires and physical factory-floor collisions through dynamic, cage-free edge safety controllers.
- Locked Capital Recovery: Resolves GST input credit blocks within days rather than months, freeing critical working capital.
- Quality Control & Output Improvement: Reduces export reject rates from 7% to under 1% with active computer vision.
- Interest Rate Reductions: Provides real-time underwriting scores to banks, securing 50–150 bps discount on commercial loans.
- Green Sourcing Incentives: Route and supplier carbon footprint indices built directly into purchasing ledgers.
- EV Battery Lifespan Extension: Edge telemetry monitors prevents extreme deep-discharge and thermal degradation patterns.
- Solar CAPEX Promotion: Automatically performs solar sizing audits and pre-fills MNRE subsidy applications for MSMEs.
| Phase | Description |
|---|---|
| Ideation | Month 1: Define the schema interfaces, write core LangGraph prompt sets, design basic FastAPI entry-points, and verify fuzzy GST reconciliation models. |
| Design | Month 2-3: Finalize hardware layouts (ESP32 CAN connections, Intel D435 placement), design PostgreSQL database schema, and mock Next.js dashboards. |
| Development | Month 4-6: Build VoltSentry ESP32 boards, customize YOLOv10 safety models on Jetson Orin Nano, and implement full agent workflows (Negotiator, Auditor). |
| Testing | Month 7-9: Edge field trials (deploy VoltSentry to 50 e-loaders and CobotSight to 3 workshops) and run automated GSTR-2B WhatsApp alert pipelines. |
| Demo Preparation | Month 10-12: Bank portal API integrations, system audits (SOC2 compliance verification), and packing final presentation decks for the MSME showcase. |
- Web/Mobile Application: Next.js 15 dashboard displaying credit ratings, supplier negotiations, battery health, and factory safety overlays.
- Prototype:
- Embedded ESP32-S3 VoltSentry CAN-bus logger.
- Jetson Orin Nano active safety loop controller with RealSense RGB-D.
- Presentation: High-fidelity project walkthrough video and system architecture slide deck.
- Source Code: GitHub repositories for the agentic backend, Next.js frontend, and C++/ROS2 edge firmware.
- Documentation: Unified database schema, hardware pinouts, wiring diagrams, and multi-agent state machines.
- Federated Fleet Diagnostics: Decentralized analytics across larger regional commercial fleets to model battery degradation curves under varied climate conditions.
- Smart Contract Escrow Sourcing: Blockchain-backed payment gates that release funds only when the SupplyAgent confirms matching GSTR-2B compliance and OCR invoice verification.
- 6-Axis Cobotic Integration: Moving from 4-axis Igus arms to 6-axis collaborative robotic integration for high-precision assembly lines.
- Operator Command: An MSME operator opens the Operionix Next.js dashboard and searches for steel tube suppliers.
- Sourcing Agent Negotiations: The Sourcing Agent finds the cheapest supplier that meets minimum ESG criteria, runs a negotiation loop, and writes a purchase ledger entry.
- Compliance Audit Trigger: The GST Auditor Agent checks GSTR-2B filings, finds a delinquent record, and automatically fires a friendly WhatsApp alert to the supplier, releasing input tax credits.
- Green Credit Recalculation: Seeing the clean compliance audit and eco-friendly procurement, the Underwriting Agent updates the company's real-time credit score, resulting in a pre-approved green credit limit upgrade at the regional cooperative bank.
- Physical Safety & Telemetry Action: While the business transactions close, the VoltSentry module monitors active EV fleets on target logistics routes. Concurrently, a worker steps too close to the sorting arm at the factory floor; the CobotSight module immediately flags this via the RealSense sensor and halts the arm, keeping the worker safe.
- Accuracy: > 98% fuzzy-match rate in invoice reconciliation; > 95% accuracy in battery state estimation.
- Response Time: < 100ms for CobotSight active safety brakes; < 2s for multi-agent negotiation decisions.
- User Satisfaction: Target Net Promoter Score (NPS) > 60 for user onboarding dashboards.
- Scalability: Handles 10,000+ edge units and 500,000 monthly transactions per database cluster.
- Cost Efficiency: Reduces telemetry cellular bandwidth costs by 80% and quality assurance labor overheads by 90%.
- Research Papers:
- Edge-based Extended Kalman Filtering for battery State of Health (SoH) and remaining useful life (RUL) estimation.
- YOLOv10: Real-Time End-to-End Object Detection on Embedded Platforms.
- APIs: GSTN GSTR-2B API, Twilio WhatsApp API, Setu Account Aggregator and GST APIs.
- Datasets: NASA Ames Battery Aging Dataset, COCO Object Detection Dataset.
- Documentation: LangGraph Stateful Multi-Agent Framework, ROS2 & MoveIt2 Trajectory Control.
- Government Frameworks & Institutional Guidelines:
- Small Industries Development Bank of India (SIDBI). Operational Guidelines for MSE Green Investment and Financing for Transformation (MSE GIFT) & End-to-End Energy Efficiency (4E) Scheme.
- Ministry of New and Renewable Energy (MNRE), Government of India. Operational Guidelines for Pradhan Mantri Kisan Urja Suraksha evam Utthaan Mahabhiyaan (PM-KUSUM) Scheme.
- Credit Guarantee Fund Trust for Micro and Small Enterprises (CGTMSE). Credit Guarantee Scheme Circulars, Eligibility and Operational Guidelines.
- Ministry of Electronics and Information Technology (MeitY), Government of India. Digital Personal Data Protection (DPDP) Act, 2023 (Gazette Notification No. 40 of 2023).
- Reserve Bank of India (RBI). Master Directions - Non-Banking Financial Company - Account Aggregator (Reserve Bank) Directions, 2016 (Updated 2023).
- Constraint: Cheap Indian electric loaders and cargo e-rickshaws do not have standard CAN bus systems or OBD-II diagnostics ports.
- Mitigation: Operionix deploys an inline Smart Shunt Retrofit Kit containing a shunt resistor (e.g., 100A/75mV), an ADC converter (ADS1115), and a digital temperature sensor array (DS18B20) wired to the ESP32-S3 host. This allows direct analog tracking of voltage drops, current throughput, and cell pack temperatures, calculating State of Health (SoH) and State of Charge (SoC) independently of vehicle controller data.
- Constraint: Small foundry and metal forging plants have extreme dust, airborne oil, and heat that cloud camera lenses, causing optical safety loops to shut down machines or fail entirely.
- Mitigation:
- Pressurized Enclosure: The depth camera is enclosed in an IP65 housing equipped with a micro-diaphragm air purge pump that continuously blows clean compressed air across the glass cover to prevent dust settling.
- Sensor-Fusion Fallback: In the event of video opacity (detected via image variance checks), the system automatically degrades gracefully from 3D camera tracking to an array of rugged, dust-insensitive industrial ultrasonic proximity sensors (e.g., MB7389 HRXL-MaxSonar-WR) mounted directly on the robotic arm, keeping safety functional at a slightly restricted velocity.
- Constraint: Cooperative and public sector banks have zero incentive or technical capacity to automatically adjust lending thresholds based on API callbacks from third-party software, fearing audit defaults.
- Mitigation: Instead of direct API control, Operionix compiles operational telemetry, GST compliance logs, and ESG parameters into a verifiable, signed PDF Report and an XML metadata package. This package is aligned with the RBI's Account Aggregator (AA) framework, which banks already use to ingest verified client profiles. MSMEs submit this report directly to get fast-tracked for pre-configured green credit schemes.
Operionix is designed to align with current Indian policies to maximize adoption subsidies and bank acceptance:
- SIDBI Green Finance Schemes: The ESG scorecards generated by the EcoVest subsystem directly map to SIDBI’s credit criteria for solar installations and energy-efficient upgrades, helping MSMEs secure loans with interest rate subventions (up to 1.5% discount).
- PM-KUSUM Scheme: Simplifies solar power adoption process by automatically calculating crop land/factory rooftop solar capacity and pre-filling Component-A/Component-C application documents for MNRE approval.
- CGTMSE (Credit Guarantee Fund Trust for Micro and Small Enterprises): Integrates transactional ledger scorecards with CGTMSE thresholds, enabling collateral-free loans by providing banks with continuous operational visibility instead of requiring land collaterals.
- DPDP Act 2023 Compliance: Ensures all MSME operational logs and personal identifiers (GSTINs, bank accounts, employee faces) undergo local anonymization and Microsoft Presidio PII scrubbing before entering cloud databases. Consent-based data-sharing ledgers allow MSMEs full control over who sees their credit and operations metadata.
📖 Click to expand Appendix A: Multi-Agent Stateful Orchestration Protocol (State Transitions)
- Idle State: Awaiting query from MSME operator containing required component specs, target price range, and minimum supplier ESG ratings.
- Marketplace Ingestion: Queries registered B2B catalogs and crawls available distributor endpoints to index match targets.
- Bidding Strategy Execution: Initiates game-theoretic price negotiations with supplier endpoints (utilizing multi-turn prompt templates).
- Verification & Logging: Checks credentials, writes final pricing, transport route proposals, and estimated carbon emission index into the Postgres store.
- Sync Trigger: Triggered by new transaction logs or automated cron jobs (e.g., 11th of every month post-GSTR-1 filing deadlines).
- Reconciliation Scan: Matches invoice fields against the GSTN GSTR-2B API output using fuzzy string and numerical alignments.
- Discrepancy Resolution: If mismatch exceeds ₹100, transitions to notification state, auto-generating a WhatsApp reminder referencing invoice ID and outstanding tax credit value.
- Resolution Logging: Transitions to resolved state when supplier updates filings, releasing locked tax credits in the active cash velocity register.
🔧 Click to expand Appendix B: Edge Hardware Schematics & Interface Specifications
The VoltSentry hardware collector acts as a robust CAN bus logger. Below is the hardware component list and electrical pin mappings:
| Component Role | Part Number / Model | Interface Protocol | Operating Voltage |
|---|---|---|---|
| System Microcontroller | ESP32-S3-WROOM-1 | System Host / WiFi / BT | 3.3V DC |
| CAN Bus Controller | MCP2515 | SPI Interface | 5.0V DC (VCC) / 3.3V logic |
| CAN Transceiver | TJA1050 | Direct CAN H/L connection | 5.0V DC |
| Cellular Modem | SIM7600G-H (4G LTE/GPS) | UART (AT Commands) | 3.8V - 4.2V (High current peaks) |
| Power Buck Regulator | LM2596 step-down | Vehicle Battery (12V/24V) to 5V | Output adjusted to 5.0V |
- Vehicle OBD-II Port (Pin 6: CAN High, Pin 14: CAN Low) connects directly to TJA1050 (CANH / CANL).
- TJA1050 Transceiver (TXD, RXD) pins connect to MCP2515 (TXCAN, RXCAN).
- MCP2515 Controller communicates with ESP32-S3 Host via SPI:
CSPin -> ESP32 GPIO 10SOPin (MISO) -> ESP32 GPIO 13SIPin (MOSI) -> ESP32 GPIO 11SCKPin -> ESP32 GPIO 12INTPin -> ESP32 GPIO 9 (for interrupt-driven CAN frame capture).
The CobotSight edge workstation relies on low-latency compute and depth imaging to run its Industry 5.0 spatial safety loops:
| Component Role | Part Number / Model | Interface Protocol | Operating Voltage |
|---|---|---|---|
| Edge Compute Unit | NVIDIA Jetson Orin Nano (8GB) | System Controller | 9V - 20V DC (15W power mode) |
| 3D RGB-D Sensor | Intel RealSense D435 | USB 3.0 Type-C | 5V (from USB host) |
| Robotic Arm | igus RBTX 4-Axis Delta / Articulated | RS485 / Modbus RTU | 24V DC |
| Actuator Stepper Driver | TB6600 High-Microstep Driver | PWM pulse + Direction | 9V - 42V DC |
- Inner Alert Boundary (Radius < 0.5m): Immediate controller interrupt. Issues E-Stop halt command to all stepper controllers.
- Warning Boundary (0.5m <= Radius <= 1.2m): Dynamic deceleration. Scales robotic arm travel velocity linearly down to 20% of max feedrate.
- Safe Zone (Radius > 1.2m): Unrestricted execution at 100% velocity.
💾 Click to expand Appendix C: Integrated Unified Database Schema
Operionix utilizes a unified relational and time-series database design to log real-time operational events alongside long-term financial profiles.
-- PostgreSQL Relational Tables (SaaS / ERP Ledger Core)
CREATE TABLE msme_profile (
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
udyam_registration_no VARCHAR(50) UNIQUE NOT NULL,
company_name VARCHAR(100) NOT NULL,
industry_type VARCHAR(50) NOT NULL,
esg_score NUMERIC(5, 2) DEFAULT 0.00,
created_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP
);
CREATE TABLE credit_limits (
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
msme_id UUID REFERENCES msme_profile(id) ON DELETE CASCADE,
approved_limit NUMERIC(15, 2) NOT NULL,
active_utilisation NUMERIC(15, 2) DEFAULT 0.00,
underwriting_score INTEGER CHECK (underwriting_score BETWEEN 300 AND 900),
interest_rate_percentage NUMERIC(4, 2) NOT NULL,
last_evaluation_date TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP
);
CREATE TABLE ledger_entries (
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
msme_id UUID REFERENCES msme_profile(id) ON DELETE CASCADE,
invoice_no VARCHAR(50) NOT NULL,
transaction_amount NUMERIC(15, 2) NOT NULL,
gst_compliance_status VARCHAR(20) CHECK (gst_compliance_status IN ('MATCHED', 'MISMATCHED', 'PENDING')),
supplier_gstin VARCHAR(15) NOT NULL,
invoice_date DATE NOT NULL,
reconciled_at TIMESTAMP WITH TIME ZONE
);
-- TimescaleDB Time-Series Table for High-Frequency EV Telemetry
CREATE TABLE battery_telemetry (
recorded_at TIMESTAMP WITH TIME ZONE NOT NULL,
vehicle_id VARCHAR(50) NOT NULL,
voltage_volts NUMERIC(5, 2) NOT NULL,
current_amps NUMERIC(6, 2) NOT NULL,
cell_temp_celsius NUMERIC(4, 1) NOT NULL,
calculated_soc_pct NUMERIC(4, 1) NOT NULL,
calculated_soh_pct NUMERIC(4, 1) NOT NULL
);
-- Convert battery_telemetry to hypertable partitioned on 'recorded_at'
SELECT create_hypertable('battery_telemetry', 'recorded_at');
-- pgvector Table for RAG-enabled Regulatory Policy Indexing
CREATE EXTENSION IF NOT EXISTS vector;
CREATE TABLE regulatory_policies (
id SERIAL PRIMARY KEY,
scheme_name VARCHAR(100) NOT NULL,
section_title VARCHAR(150) NOT NULL,
policy_chunk_text TEXT NOT NULL,
chunk_embedding vector(1536) -- Compatible with standard openai ada-002 embeddings
);📈 Click to expand Appendix D: Commercialization & Scalability
- Modular SaaS Licensing: A pay-as-you-grow model. MSMEs can start with the free software module (GST audits) and add hardware modules (VoltSentry OBD links, CobotSight kits) as operational requirements expand.
- Cooperative Bank SaaS Integrations: Subscriptions sold directly to regional cooperative and rural banks, who distribute the platform to their MSME clients to reduce credit defaults and de-risk portfolios.
- Certified Partner Network: Local automation technicians trained as certified Operionix system integrators, providing localized physical assembly and calibration services to MSMEs across industrial hubs.